Bladder cancer (BLC) is one of the most common urological malignancies worldwide and is characterized by high biological heterogeneity, extremely variable recurrence rates and often unpredictable therapeutic responses. Despite significant advances in precision oncology, the clinical application of personalized strategies for BLC remains limited, mainly due to the lack of preclinical models capable of faithfully reproducing the structural, molecular and microenvironmental complexity of human tumors. Traditional two-dimensional models fail to maintain the three-dimensional architecture and the heterotypic interactions that define the tumor microenvironment (TME), while more advanced systems such as patient-derived xenografts (PDX), although biologically faithful, are costly, require prolonged development times and are not suitable for rapid clinical decision-making. In this context, patient-derived organoids (PDOs) are emerging as highly valuable tools capable of integrating both molecular and functional aspects of the disease. The main objective of this doctoral thesis was to develop, characterize and evaluate the clinical applicability of BLC PDOs as reliable preclinical models able to reproduce the complexity of bladder tumors and supporting personalized therapeutic strategies. To this end, a cohort of patients treated at the IRCCS Regina Elena National Cancer Institute (IRE) in Rome was analysed, and tumor samples were obtained by transurethral resection of bladder tumor (TURBT) or radical cystectomy. For each patient, samples were collected from two spatially distinct tumor regions, central (TC) and peripheral (TP), with the aim of investigating intratumoral heterogeneity and its impact on treatment responses. The first part of the work focused on optimizing PDO culture conditions. Through a systematic comparison of matrices and culture media, it was possible to achieve a 60% success rate, significantly higher than values reported in many published studies. In particular, the use of Ultra-Low Attachment (ULA) plates and an enriched culture medium (Growth Medium 3) enabled more efficient PDO generation compared with matrix-based systems such as Geltrex. A selection of PDOs was subsequently characterized through immunohistochemistry (IHC), Whole Exome Sequencing (WES) and RNA Sequencing (RNA-seq), showing a high degree of molecular and transcriptional concordance with the original tissues. WES analysis revealed the preservation of typical BLC driver mutations, such as FGFR3, TP53, ERBB2 and KDM6A, while transcriptomic analysis confirmed the comparability of major oncogenic pathways (MYC, PI3K/AKT/mTOR, WNT, p53), highlighting the ability of PDOs to faithfully maintain the molecular identity of the tumor. The investigation of intratumoral heterogeneity represented a central focus of this thesis. PDOs generated from TC and TP regions displayed significant mutational differences, including region-specific loss of mutations and variations in Tumor Mutational Burden (TMB). Although transcriptomic profiles were largely comparable, genomic and functional data revealed distinct treatment responses between the two regions. For example, in patient BLC17, both TC and TP PDOs harboured the same FGFR3 mutation, only the PDO derived from the central region showed sensitivity to the FGFR inhibitor Erdafitinib, suggesting that non-genetic factors, such as cellular plasticity, TME-related cues or transient transcriptional states, may modulate drug response. Similarly, intra-patient differences were observed in responses to Cisplatin and Gemcitabine, confirming that the spatial variability of bladder tumors can significantly influence therapeutic decisions. An innovative aspect of the project involved assessing the preservation of the tumor microenvironment in PDOs. Through ssGSEA analyses and immune deconvolution, it was shown that organoids can retain a substantial portion of the stromal and immune components of the tumor, although with adaptations typical of in vitro culture. Transcriptional signatures associated with fibroblasts, extracellular matrix and angiogenesis were partially preserved, while certain immune populations, such as dendritic cells, monocytes and M2 macrophages, were reduced. Nonetheless, PDOs maintained sufficient levels of CD45+ and PD-1+ cells to allow functional immunotherapy testing. Based on these observations, an analysis was conducted to evaluate PDO responses to Nivolumab, a monoclonal antibody targeting PD-1, both as monotherapy and in combination with standard chemotherapy (Cisplatin + Gemcitabine). The results revealed that organoids derived from tumors with high immune infiltration exhibited a marked reduction in viability when treated with the chemo-immunotherapy combination, whereas those with low infiltration remained refractory. Nivolumab alone did not exert any cytotoxic effect, confirming that immunotherapy efficacy in PDOs depends on the functional presence of immune cells. This evidence suggests that early-passage PDOs represent a suitable model for studying immunotherapy responses, although the useful experimental window is limited by the progressive loss of immune components during prolonged cultures. Overall, this thesis demonstrates that PDOs are a robust and clinically relevant preclinical model for studying BLC. They allow for the faithful reproduction of the molecular and phenotypic features of the original tumors, to capture intratumoral heterogeneity, to perform functional drug screening within timelines compatible with clinical practice and to personalize the evaluation of responses to standard treatments, targeted therapies and immunotherapies, as well as to analyze, at least at early passages, critical components of the TME such as immune infiltrates and stromal transcriptional signatures. These findings strengthen the role of PDOs as promising platforms for the integration of precision medicine in BLC and support their future use in real-time clinical workflows, co-clinical studies and therapeutic stratification programs based on functional evidence.
Frascolla, C. (2026). Preservation of Immune Features in early passage Bladder Cancer Organoids Enables Assessment of Immunotherapy Responses.
Preservation of Immune Features in early passage Bladder Cancer Organoids Enables Assessment of Immunotherapy Responses
Carlotta Frascolla
2026-04-27
Abstract
Bladder cancer (BLC) is one of the most common urological malignancies worldwide and is characterized by high biological heterogeneity, extremely variable recurrence rates and often unpredictable therapeutic responses. Despite significant advances in precision oncology, the clinical application of personalized strategies for BLC remains limited, mainly due to the lack of preclinical models capable of faithfully reproducing the structural, molecular and microenvironmental complexity of human tumors. Traditional two-dimensional models fail to maintain the three-dimensional architecture and the heterotypic interactions that define the tumor microenvironment (TME), while more advanced systems such as patient-derived xenografts (PDX), although biologically faithful, are costly, require prolonged development times and are not suitable for rapid clinical decision-making. In this context, patient-derived organoids (PDOs) are emerging as highly valuable tools capable of integrating both molecular and functional aspects of the disease. The main objective of this doctoral thesis was to develop, characterize and evaluate the clinical applicability of BLC PDOs as reliable preclinical models able to reproduce the complexity of bladder tumors and supporting personalized therapeutic strategies. To this end, a cohort of patients treated at the IRCCS Regina Elena National Cancer Institute (IRE) in Rome was analysed, and tumor samples were obtained by transurethral resection of bladder tumor (TURBT) or radical cystectomy. For each patient, samples were collected from two spatially distinct tumor regions, central (TC) and peripheral (TP), with the aim of investigating intratumoral heterogeneity and its impact on treatment responses. The first part of the work focused on optimizing PDO culture conditions. Through a systematic comparison of matrices and culture media, it was possible to achieve a 60% success rate, significantly higher than values reported in many published studies. In particular, the use of Ultra-Low Attachment (ULA) plates and an enriched culture medium (Growth Medium 3) enabled more efficient PDO generation compared with matrix-based systems such as Geltrex. A selection of PDOs was subsequently characterized through immunohistochemistry (IHC), Whole Exome Sequencing (WES) and RNA Sequencing (RNA-seq), showing a high degree of molecular and transcriptional concordance with the original tissues. WES analysis revealed the preservation of typical BLC driver mutations, such as FGFR3, TP53, ERBB2 and KDM6A, while transcriptomic analysis confirmed the comparability of major oncogenic pathways (MYC, PI3K/AKT/mTOR, WNT, p53), highlighting the ability of PDOs to faithfully maintain the molecular identity of the tumor. The investigation of intratumoral heterogeneity represented a central focus of this thesis. PDOs generated from TC and TP regions displayed significant mutational differences, including region-specific loss of mutations and variations in Tumor Mutational Burden (TMB). Although transcriptomic profiles were largely comparable, genomic and functional data revealed distinct treatment responses between the two regions. For example, in patient BLC17, both TC and TP PDOs harboured the same FGFR3 mutation, only the PDO derived from the central region showed sensitivity to the FGFR inhibitor Erdafitinib, suggesting that non-genetic factors, such as cellular plasticity, TME-related cues or transient transcriptional states, may modulate drug response. Similarly, intra-patient differences were observed in responses to Cisplatin and Gemcitabine, confirming that the spatial variability of bladder tumors can significantly influence therapeutic decisions. An innovative aspect of the project involved assessing the preservation of the tumor microenvironment in PDOs. Through ssGSEA analyses and immune deconvolution, it was shown that organoids can retain a substantial portion of the stromal and immune components of the tumor, although with adaptations typical of in vitro culture. Transcriptional signatures associated with fibroblasts, extracellular matrix and angiogenesis were partially preserved, while certain immune populations, such as dendritic cells, monocytes and M2 macrophages, were reduced. Nonetheless, PDOs maintained sufficient levels of CD45+ and PD-1+ cells to allow functional immunotherapy testing. Based on these observations, an analysis was conducted to evaluate PDO responses to Nivolumab, a monoclonal antibody targeting PD-1, both as monotherapy and in combination with standard chemotherapy (Cisplatin + Gemcitabine). The results revealed that organoids derived from tumors with high immune infiltration exhibited a marked reduction in viability when treated with the chemo-immunotherapy combination, whereas those with low infiltration remained refractory. Nivolumab alone did not exert any cytotoxic effect, confirming that immunotherapy efficacy in PDOs depends on the functional presence of immune cells. This evidence suggests that early-passage PDOs represent a suitable model for studying immunotherapy responses, although the useful experimental window is limited by the progressive loss of immune components during prolonged cultures. Overall, this thesis demonstrates that PDOs are a robust and clinically relevant preclinical model for studying BLC. They allow for the faithful reproduction of the molecular and phenotypic features of the original tumors, to capture intratumoral heterogeneity, to perform functional drug screening within timelines compatible with clinical practice and to personalize the evaluation of responses to standard treatments, targeted therapies and immunotherapies, as well as to analyze, at least at early passages, critical components of the TME such as immune infiltrates and stromal transcriptional signatures. These findings strengthen the role of PDOs as promising platforms for the integration of precision medicine in BLC and support their future use in real-time clinical workflows, co-clinical studies and therapeutic stratification programs based on functional evidence.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


